add version and capabilities, some cleanup still required, but code works
git-svn-id: file:///svn/mysql/tokudb-engine/src@4929 c7de825b-a66e-492c-adef-691d508d4ae1
remove update_status, at moment, there is nothing that needs updating when
DB is going away
git-svn-id: file:///svn/mysql/tokudb-engine/src@4926 c7de825b-a66e-492c-adef-691d508d4ae1
as of now, no info is stored in status.tokudb because the old info
that was stored is not needed anymore. Info will be stored soon.
git-svn-id: file:///svn/mysql/tokudb-engine/src@4920 c7de825b-a66e-492c-adef-691d508d4ae1
remove unused variables share->rows and share->org_rows that were useed
to estimate row count. row count is now estimated with key_range64
git-svn-id: file:///svn/mysql/tokudb-engine/src@4919 c7de825b-a66e-492c-adef-691d508d4ae1
make row count displayed in show table status better by
using key_range64 to estimate stats.records instead of keeping a running tally
git-svn-id: file:///svn/mysql/tokudb-engine/src@4918 c7de825b-a66e-492c-adef-691d508d4ae1
temporary change to test if just doing a memcpy will work for unpacking
a Field_new_decimal
real change is to modify Field_new_decimal, but by checking this in, we can
quickly get regression results
git-svn-id: file:///svn/mysql/tokudb-engine/src@4904 c7de825b-a66e-492c-adef-691d508d4ae1
just unpacking key has problems when the key is a substring of a column. MySQL still reports that it is a covering index, must back out previous change
git-svn-id: file:///svn/mysql/tokudb-engine/src@4782 c7de825b-a66e-492c-adef-691d508d4ae1
dont unpack entire row when primary key is used as covering index
git-svn-id: file:///svn/mysql/tokudb-engine/src@4770 c7de825b-a66e-492c-adef-691d508d4ae1
back out change for 4591, need more testing before it can go in
reopening ticket
git-svn-id: file:///svn/mysql/tokudb-engine/src@4634 c7de825b-a66e-492c-adef-691d508d4ae1
fix buffer overwrite error. Bug has been around, but just recently exposed
by checkin 4590
git-svn-id: file:///svn/mysql/tokudb-engine/src@4616 c7de825b-a66e-492c-adef-691d508d4ae1
implement read_time to take into account that the primary_key is clustered
git-svn-id: file:///svn/mysql/tokudb-engine/src@4595 c7de825b-a66e-492c-adef-691d508d4ae1
pre_acquire table locks from lock tree when mysql is doing a lock table
git-svn-id: file:///svn/mysql/tokudb-engine/src@4591 c7de825b-a66e-492c-adef-691d508d4ae1
rename lock_grabbed to range_lock_grabbed to help distinguish from
cases where table lock is grabbed
git-svn-id: file:///svn/mysql/tokudb-engine/src@4589 c7de825b-a66e-492c-adef-691d508d4ae1
fix index_prev in case we don't have lock
it was a typo before
git-svn-id: file:///svn/mysql/tokudb-engine/src@4516 c7de825b-a66e-492c-adef-691d508d4ae1
make handlerton not use a cursor that has had a failed operation
this will be useful for lightweight cursors
git-svn-id: file:///svn/mysql/tokudb-engine/src@4455 c7de825b-a66e-492c-adef-691d508d4ae1
slightly restructure some functions to allow handling of lighter weight cursors
git-svn-id: file:///svn/mysql/tokudb-engine/src@4439 c7de825b-a66e-492c-adef-691d508d4ae1
used to disable covering indexes for some keys with strings for wrong reasons
not anymore
git-svn-id: file:///svn/mysql/tokudb-engine/src@4386 c7de825b-a66e-492c-adef-691d508d4ae1
add instances of read_range_first and read_range_next in ha_tokudb
currently just call parent class version
git-svn-id: file:///svn/mysql/tokudb-engine/src@4385 c7de825b-a66e-492c-adef-691d508d4ae1
make add_index faster by making inserts into secondary table all be part
of on transaction
git-svn-id: file:///svn/mysql/tokudb-engine/src@4287 c7de825b-a66e-492c-adef-691d508d4ae1
check to see if keys that are added have same name as existing keys.
if so, gracefully return error
git-svn-id: file:///svn/mysql/tokudb-engine/src@4250 c7de825b-a66e-492c-adef-691d508d4ae1
table->s->max_key_length does not take into account new keys that are being added, and therefore is not suitable as a buffer size
git-svn-id: file:///svn/mysql/tokudb-engine/src@4167 c7de825b-a66e-492c-adef-691d508d4ae1
